Evan Simpson wrote:
It was asserted, last time this came up, that this is compliant with the
ISO8601 spec, but my own research shows this to be false. If someone
can point me to a truly authoritative source that supports the current
behavior, I would appreciate it, but that would not change the
This has been brought up and (rather cavalierly) dismissed before. I
challenge anyone to justify the following DateTime behavior:
>>> DateTime('2005-01-01').strftime('%Y-%m-%d')
'2004-12-31'
DateTimes are actually timestamps, and the current implementation
assumes that a date in ISO8601 format w